OpenSHMEM and Related Technologies. OpenSHMEM in the Era of Extreme Heterogeneity

This book constitutes the proceedings of the 5th OpenSHMEM Workshop, held in Baltimore, MD, USA, in August 2018.

The 14 full papers presented in this book were carefully reviewed and selected for inclusion in this volume. The papers discuss a variety of ideas for extending the OpenSHMEM specification and discuss a variety of concepts, including interesting use of OpenSHMEM in HOOVER a distributed, flexible, and scalable streaming graph processor and scaling OpenSHMEM to handle massively parallel processor arrays. The papers are organized in the following topical sections: OpenSHMEM library extensions and implementations; OpenSHMEM use and applications; and OpenSHMEM simulators, tools, and benchmarks.

Inhalt

OpenSHMEM Library Extensions and Implementations.- OpenSHMEM Sets and Groups: An Approach to Worksharing and Memory Management.- Design and Optimization of OpenSHMEM 1.4 for the Intel® Omni-Path Fabric 100 Series.- Introducing Cray OpenSHMEMX - A Modular Multi-Communication Layer OpenSHMEM Implementation .- An Initial Implementation of Libfabric Conduit for OpenSHMEM-X.- The OpenFAM API: A Programming Model for Disaggregated Persistent Memory.- SHCOLL - A Standalone Implementation of OpenSHMEM-style Collectives API.- OpenSHMEM Use and Applications.- HOOVER: Distributed, Flexible, and Scalable Streaming Graph Processing on OpenSHMEM.- Tumbling Down the GraphBLAS Rabbit Hole with SHMEM.- Scaling OpenSHMEM for Massively Parallel Processor Arrays.- Designing High-Performance In-Memory Key-Value Operations with Persistent GPU Kernels and OpenSHMEM.- OpenSHMEM Simulators, Tools, and Benchmarks.- Towards Lightweight and Scalable Simulation of Large-Scale OpenSHMEM Applications.- Tracking Memory Usage in OpenSHMEM Runtimes with the TAU Performance System.- Lightweight Instrumentation and Analysis using OpenSHMEM Performance Counters.- Oak Ridge OpenSHMEM Benchmark Suite.